I chair UNC’s Educational Policy Committee, and we are in the process of seeking some new policy initiatives to address grading. A reporter for the Daily Tar Heel asked me a while ago why I am such a grading “hawk”, meaning that I worry about grading problems (more on the identification of these problems below the break). The reason for his question is that I am a relatively humanities-oriented scholar in a department and discipline not exactly known for rigorous grading policies. Below the break I’ll discuss what I see as the problems, possible solutions, UNC’s current status with regard to these solutions, and why I care so much about them. Warning: this is a long and somewhat rambling post.
First, some definitions. I will refer to grade inflation as the general rise in grades over time, even though this leaves open the question of whether the “product” (quality of student work) is getting better, in which case the inflation metaphor is inexact at best. Grade compression, on the other hand, is a result of the combination of grade inflation raising the floor and the nature of the grading scale including a ceiling, such that all variation in student achievement must be represented in a narrowing window (since we don’t, as of yet, have AA, AAA, etc. grades). Finally, systematic grading inequality refers to the fact that grades received by students vary in far-from-random fashion across disciplines and (crucially) instructors within disciplines. Each of these is a distinct problem, even though they’re of course closely interrelated. Each raises a different set of thorny issues. And each is amenable to a different set of policy interventions, though again, there is much overlap.
What are the harms?
Lots of people agree that grades are rising, are unequal across departments and instructors, and are becoming more compressed, but ask who is harmed. After all, just about everyone is getting higher grades; professors are getting fewer complaints; graduate instructors and TAs have less grading hassles to worry about. It’s a win-win solution, right?
Wrong. Here’s a list of reasons to care about this set of problems:
- Accuracy. We have a responsibility to provide accurate information to
students and the public we serve
- Reduction in value. An A at Carolina (or wherever) means less if it is the most common grade
- Incentives. It is very difficult to recognize and reward outstanding
- Perverse incentives. Inappropriate incentives to select some courses or majors
- Unfairness. Cross-department comparisons and rankings are invalid and
The pervasive use of GPA as a mode of cross-class comparison and student ranking is a case of reactivity: students have altered behavior, in some cases radically, in the service of gaining or maintaining a GPA that reflects their aspirations. That means, ceteris paribus, that students are selecting into courses and potentially even majors that are not the best intellectual homes for them because of a statistical artifact.
At UNC, since 1995 the proportion of students eligible for the Dean’s List has climbed from 25% to 40%. The likelihood is that it will climb further soon, except that my committee has been asked to review the situation. We will be approaching the Faculty Council with the problem and a set of proposals for handling the Dean’s List in specific, but this is clearly a symptom that will not really be fixed unless and until we handle grades in general. Consider, for example, if we were to set 25% as a goal and just set GPA targets based on that goal (one of the proposals). This serves to magnify inequality based on instructor and department and, in turn, increase perverse incentives on students and faculty alike, to the extent that Dean’s List is a motivating distinction.
What to do?
But I digress. Here’s a (fairly) comprehensive list of possible policy responses to the problems:
- Separate evaluation of student performance from teaching (e.g., Swarthmore’s honors program). Professors concentrate on teaching; a distinct mechanism is devised to assess achievement. If done right, this is an intellectually attractive but financially disastrous plan. If done wrong, it is financially feasible but intellectually disastrous (think No Child Left Behind at the college level).
- Limit instructor and/or department grades through rationing (e.g., Princeton) or defined averaging (e.g., Wellesley). Recognize that grades are in some sense zero-sum, so treat them as scarce goods on the “front end.” This is appealing in that it aims actually to fix the ontological assessment problem instead of simply fiddling around. But it has serious implications vis-a-vis intellectual freedom and teaching philosophy that ought to be a major concern philosophically. Here at UNC, it was clear at my presentation this fall that this class of options was unacceptable among the Faculty Council.
- Report context information for each grade on the transcript (e.g., Indiana, Cornell). Thus prospective employers, graduate schools, etc., could evaluate the relative difficulty involved in achieving each grade on the transcript. However, aggregate comparisons among students would remain invalid.
- Provide an adjusted measure of accomplishment for use in reporting relative performance and rankings. This is what we tried to do a few years ago with the Achievement Index, which still leaves a bitter taste in the mouths of some around UNC. Ironically, though, it was probably the most popular option during the Faculty Council discussion in which I outlined these options. Essentially this would involve pulling in information about the relative difficulty of achieving a student’s grade mix to develop a valid aggregate comparison of student achievement for use in Dean’s List, Distinction, and other comparative tasks. (GPA would probably remain alongside it for comparability, even though it is essentially meaningless as a measure of overall performance.)
- Prohibit comparison of students’ accomplishment across departments and instructors. Since we know unequivocally that grades without any adjustment are an invalid way of comparing students’ achievement, the University could choose to acknowledge this by simply saying we have no way of evaluating our students’ overall achievement. While of course we cannot keep third-party consumers from doing what they want with reported grades, we could abolish Dean’s List, Distinction, etc.; prohibit all University units from using grades to compare student achievement for fellowships, honors, etc.; and prohibit the Registrar from reporting any aggregate information such as GPA.
- Ongoing, University-wide discussion and deliberation process on grading meaning and philosophy (e.g., Seton Hall). Hey, everybody likes the idea of ongoing deliberation. But we would need to do much, much more to establish a kind of Durkheimian moral order in which the grade is a sacred instrument not to be messed with.
- Watch and wait (do nothing now). Always a popular one. The big question: if not now, when? What would constitute a crisis sufficient to warrant intervention if the situation we now find ourselves in isn’t it?
An additional concern is that this is a national trend, not a single university’s problem. Furthermore, our graduates will find themselves in competition for graduate school admissions and jobs with graduates of other universities who are not, or may not be, changing practices. To mitigate this concern, the Faculty Council has asked us to build bridges with faculty senates at other universities interested in addressing these issues. While I agree in principle that this is a good idea, I am less concerned with students’ prospects because I think UNC making any bold move on this issue will make enough headlines that grade “consumers” will know what we’ve done.
UNC’s Status and Next Steps
At the January meeting of the Faculty Council, we have been asked to present the Dean’s List problem and possible solutions. The options we plan to present will be:
- Each spring, set a GPA “bar” for the following academic year targeted at making roughly 25% of students eligible for Dean’s List.
- Reduce the bar to a 2.0 GPA and publicize that fact, thereby emphasizing that the Dean’s List is meaningless; such a low GPA bar will also reduce the perverse incentives to instructors and students.
- Abolish the Dean’s List altogether.
In April we have been asked to present a comprehensive policy proposal on grading to Faculty Council. That will be the real “zinger.” I hope we can present a proposal for some sort of valid comparison statistic that will begin to reduce the inequality problems at least!
Why do I care so much?
Sociology is too often seen as an “Easy A” discipline; I’ve had far too many students tell me to my face that they’re taking my class in order to get an easy A. Two students asked me this semester if I would sign special forms to allow them to drop my class after the final exam because they didn’t like the grade they expected to get. I feel that the current grading regime is bad for the intellectual quality of the education we can offer our students. It’s bad for the “good” students because they can’t be rewarded for outstanding work; it’s bad for “bad” students because they aren’t motivated to do better. While this comes across as being a conservative hard-ass, and I’m the first to admit that I hate having to talk to students who are disappointed by their grades, I think it’s our responsibility as serious scholars and educators to maintain and increase the quality of our education–which includes seriously, carefully, and judiciously evaluating students’ real achievement in our classes.